The database I would go to first is the most used database in the world by a long way - SQLite3. It runs on everything, is designed for up to 140TBytes. If it is run in RAM, it is the fastest major SQL database in the world up to sizes of a few TBytes (using a platform like Bedrock -
https://bedrockdb.com- to handle all the DB synching to hard drives in a robust, professional way). Microsoft Access is a pathetic toy compared to SQLite3.
SQLite3 runs fine on Windows, but Azure supports Linux big time. That means you can easily run Bedrock. If you must run Windows, it now has a Linux subsystem you can enable, and so you can have subsystems of Ubuntu, Debian, Centos, etc running on Windows if you insist. It is just another option that is available. Microsoft is actually one of the Platinum members of the Linux Foundation.
There are some SQL GUI front ends that work fine with SQLite3 - you could even use the Access one or the Base interface in LibreOffice, but the thing about databases is the GUI tools never take you that far. For a decent database, you usually need to start coding. Without the coding, the database structure is usually weak, and it only works as long as users enter data in an exact specified way. When a database is exposed to a test where anyone is allowed to use it, it is amazing the genius of the average person to discover ways to incorrectly use a database in a way that that breaks it.
You can use spreadsheets as DB frontends if you want.
I haven't tried it, but there are commercial database agnostic development tools such as
https://www.sqlmaestro.com.
The actual SQLite3 code is tiny which is why it is so popular - under 1MBytes for 64 bit Windows. A tiny size has from the start been a primary design mission.
Richard